Rain can reduce a coffee shop’s demand while also delaying deliveries. Reacting only to a delayed shipment can cause unnecessary spending; waiting too long can cause a stockout.
RainCheck Restock combines inventory, demand assumptions, incoming orders and supplier quotes to decide whether to wait or purchase emergency supplies.
Airbyte’s Google Sheets connector, run through PyAirbyte, reads four business-data streams. Our backend loads the records into SingleStore, queries them for evaluation, and stores immutable snapshots, decisions, purchase-order drafts and audit events.
A deterministic engine calculates shortages, adds a safety buffer, rounds to whole packs, and compares supplier availability, delivery time and total cost including shipping. Claude uses read-only tools to explain the results.
In our demo, 300 cups cover five days of assumed rainy-day demand. A shipment arriving in four days requires no emergency purchase. Changing its ETA to six days triggers a recommendation to buy 200 cups for $26 from the cheapest supplier meeting the arrival constraint. Users can create a draft without placing a real order.
Sponsor integrations are live; business data, weather effects and delivery scenarios are synthetic. Weather/news ingestion and learning from historical outcomes are future work.